For health and life sciences organizations, the AI conversation is changing.
The question is no longer simply where AI can be applied. Leaders now need to determine which investments can deliver measurable value at enterprise scale while managing cost, risk and accountability.
This creates an important executive challenge: How do you scale AI without allowing its economics to scale unchecked?
Part of the answer lies in understanding and governing AI consumption.
AI introduces a different cost equation
Health and life sciences organizations face rising patient demand, talent scarcity, tighter budgets, pressure to speed up research and development (R&D) and changing regulatory requirements. At the same time, many are moving AI from pilots into production. Our 2026 Voice of Our Clients research shows that leaders are prioritizing high-value AI use cases and strengthening governance as they pursue measurable business outcomes.
While traditional technology investments can often be planned around relatively familiar cost structures, Generative AI (GenAI) changes the equation. Consumption varies based on how models are used, which models are selected and how agents and applications are designed.
Tokens, the units processed by many GenAI models, are one component of that consumption. Individually, their cost may appear small. At enterprise scale, however, leaders need to know where consumption occurs and whether it creates sufficient value.
This is particularly relevant in health and life sciences, where AI applications can span clinical and administrative workflows, claims, R&D and other information-intensive processes.
Without effective governance, organizations can struggle to understand AI spending, allocate costs to the appropriate business solutions and products and establish accountability. Costs can also escape scrutiny during development and testing, making overall AI spending harder to forecast.
AI cost management therefore cannot begin after deployment. Organizations need to build it into their operating model before starting pilots.
Moving from cost visibility to accountability
Effective AI cost governance connects technical consumption to business ownership.
Leaders need to understand which AI resources teams consume, who consumes them, which use cases they support and what outcomes those use cases are expected to achieve.
A practical governance model establishes clear ownership, cost allocation principles and accountability for AI resources. Metrics and reporting can then help leaders monitor consumption and make informed decisions.
Similar challenges around cost governance emerged during the early adoption of cloud services. As organizations scaled their cloud environments, many saw costs grow rapidly without the governance and visibility needed to manage them effectively. Cloud FinOps emerged in response, bringing financial accountability, governance and cost optimization into cloud operating models.
As enterprise AI approaches a similar inflection point, the lesson from cloud is clear: organizations should establish the financial and governance disciplines for AI early, before adoption outpaces their ability to understand, control, and optimize its costs.
For generative AI, token consumption can be part of that picture. Tracking tokens at the appropriate level can help teams identify inefficient usage and assess design choices. Leaders can then evaluate that consumption against the business metrics that determine whether a use case warrants further investment.
The objective is not to minimize token use. It is to optimize AI consumption relative to the value it creates.
Connecting governance with architecture
Cost governance is not solely a finance responsibility. Architecture decisions also influence how efficiently organizations consume AI resources.
Cloud architectures that do not support effective cost management, limited use of capabilities such as autoscaling and automatic shutdown and the application of on-premises practices to cloud environments can weaken cost control. As organizations scale AI, technical, financial and business decisions need to support the same objectives.
Data governance is equally important. According to our 2026 VOC research, 54% of the health and life sciences business and IT leaders interviewed report having a holistic data strategy across their internal enterprise, while only 17% report a centralized approach to internal data governance. Across external ecosystems of partners and suppliers, those figures fall to 32% and 14%, respectively.
This gap matters because enterprise AI depends on trusted data and governance that extends beyond individual experiments.
Making economics part of enterprise AI governance
For health and life sciences executives, AI governance needs to address more than security, responsibility and compliance. It also needs to determine whether AI is economically sustainable.
Three actions can help:
- Establish ownership before scaling: Define who owns each AI resource and use case, who is accountable for consumption and how costs will be allocated.
- Measure consumption alongside outcomes: Monitor relevant cost and usage metrics, including token consumption where appropriate, and connect them to business measures. A lower AI bill is not necessarily a better result if it reduces the value of the application.
- Build continuous optimization into governance: AI services, pricing and purchasing options continue to change. Establish an ongoing process for reviewing costs, architecture choices and consumption patterns rather than relying on a one-time business case.
For more guidance on this topic, I invite you to read recent blogs from our global AI team: The CFO's AI token economy playbook: Why agentic cost management is becoming a boardroom issue and The AI token economy: Managing the hidden costs of agentic AI.
From AI experimentation to sustainable value
Health and life sciences organizations are moving from AI experimentation toward operational execution. Enterprise-ready AI will require modern foundations, meaningful metrics and human-led governance embedded in individual use cases.
Economic governance needs to be part of that approach. For years, organizations have relied on established governance structures, such as project management offices (PMOs) to oversee programs and projects, and enterprise architecture boards to guide technology and architecture decisions. Rather than creating entirely new governance layers for AI, organizations can build on these proven structures, embedding relevant AI considerations and decision rights into existing processes. This approach can strengthen AI oversight while minimizing unnecessary complexity and duplication.
As AI adoption expands, leaders need to understand both what their organizations consume and what they gain in return. Token consumption provides one useful measure, but the broader objective is to make informed investment decisions that connect AI costs with measurable outcomes.
For executives, the question is no longer simply, “How much AI are we using?” A more useful question is: “Are we governing AI consumption in a way that allows us to scale its value?”
If your organization is considering how to build economic governance into its AI strategy, I welcome the opportunity to compare perspectives. Reach out to me to discuss how health and life sciences leaders can plan for AI costs, accountability and value as they move from experimentation to enterprise scale.
Read Part 1 of this series, Digital reengineering in health & life sciences: From transformation ambition to measurable results, to explore why scaling AI with discipline is essential to turning transformation ambition into measurable business value.
Back to top